Artificial intelligence, big tech and the talent base — overview and summary
Read the PDF edition (4 pages)
A financial centre has always depended on the people who work in it. The merchant bankers of the nineteenth century, the securities traders of the post-war decades and the wealth managers who turned Zurich into a global address for private fortunes each brought a specific combination of knowledge, language and trust. Since the second decade of the twenty-first century, a new group has joined them: machine learning researchers, data engineers and software developers, many of whom arrived in the city because of ETH Zurich and stayed because Google, Microsoft, Apple, Meta and, more recently, OpenAI and Anthropic opened laboratories within walking distance of the main railway station. The chapter examines how this concentration of technical talent came about and what it means for the banks, insurers, asset managers and law firms that make up the financial centre.
The subject belongs in a book on Zurich as a financial centre for two reasons. Artificial intelligence is changing the daily work of every regulated institution, from the screening of new clients and the monitoring of payments to underwriting, portfolio construction and the drafting of legal documents. The second reason is regulatory. Switzerland has chosen a path of its own, which combines the ratification of the Council of Europe’s AI Convention with targeted amendments to sectoral law, while the Swiss Financial Market Supervisory Authority (FINMA) has set out its expectations for the governance of AI in supervisory guidance. Institutions in Zurich must reconcile these Swiss rules with the European Union’s AI Act whenever their systems reach clients or markets in the Union, which is the case for most of them.
The chapter proceeds from the academic foundation to the market and then to the law. It first describes the research institutions, above all ETH Zurich and the University of Zurich, and the computing infrastructure on which Swiss research relies. It then turns to the large technology companies, the start-up scene and its financing, and the conditions under which foreign specialists can live and work in Zurich. The second half deals with AI in banking, insurance and compliance, the supervisory expectations of FINMA, the Swiss approach to AI regulation, data protection, outsourcing and cloud computing, operational resilience and cyber risk, and finally the use of AI in legal practice under the rules on professional secrecy. Fintech licensing and the DLT legislation are treated in chapter J.I, tokenisation and market infrastructure for crypto-assets and Digital Assets in chapter J.II, and the basic anti-money laundering framework in chapter I.III, so these topics appear here only where they intersect with artificial intelligence.
Conclusion
The history traced in this chapter is short compared with the centuries of banking and insurance described in the earlier parts of the book, yet it follows a familiar pattern. Zurich became a centre of artificial intelligence for the same reasons that it became a financial centre. A strong educational institution trained people of high quality, open borders and a stable legal order allowed foreign firms to settle, and a pleasant city persuaded specialists to stay. ETH Zurich plays for AI the role that the commercial and legal knowledge of the old merchant houses played for banking. Google’s decision of 2004 to build its engineering centre in Zurich functioned in the way that the founding of the large banks did in the nineteenth century, because it created a pool of experienced people from which later arrivals could hire. OpenAI and Anthropic came in 2024 and 2025 because the researchers they wanted already lived in the city.
For the financial centre, this concentration of talent is an asset that no competing European centre can easily reproduce. Banks, insurers and asset managers in Zurich can recruit data scientists who were trained at ETH, gained experience at Google or Microsoft and are familiar with the requirements of regulated industries. They can work with start-ups such as Unique or DeepJudge that were built for their needs, and they can test sovereign models such as Apertus on Swiss infrastructure. The constraints are equally clear. Housing is scarce and expensive, permits for third-country nationals are limited, and the frontier laboratories can pay salaries that regulated institutions cannot match. The domestic venture capital market remains small, so that large growth rounds usually require international investors.
The legal framework for AI in finance is in a transitional state. FINMA has chosen to treat AI within its existing supervisory categories, and Guidance 08/2024 sets out expectations on governance, inventory, risk classification, data quality, testing, documentation and explainability that are now the standard against which institutions are measured. The Federal Council’s decision of 12 February 2025 to ratify the Council of Europe Convention and to amend sectoral law has not yet produced a consultation draft. The Federal Council has promised one by the end of 2026. In the meantime, the revised FADP, the outsourcing rules of Circular 2018/3, banking secrecy, the resilience requirements of Circular 2023/1 and the new cyber reporting duties form a dense body of rules that apply to AI without mentioning it. The EU AI Act adds a further layer for every institution whose AI outputs are used in the Union, with the high-risk obligations now scheduled for December 2027.
Human capital is therefore the decisive location factor in this field. Models can be downloaded and computing capacity can be rented, but the ability to deploy AI responsibly in a bank, an insurer or a law firm depends on people who understand the technology, the business and the law at the same time. Zurich’s advantage lies in the density of such people. Its future as a financial centre will depend to a considerable degree on whether the city, the canton and the Confederation can keep the conditions that brought them here, from the quality of ETH and the openness of the labour market to the availability of homes in which they can live.
In five sentences
1. Zurich became a European centre of artificial intelligence because ETH Zurich, the ETH AI Center founded in 2020 and the national infrastructure of the Swiss AI Initiative, the Alps supercomputer and the open Apertus model created a research base that attracted Google, Microsoft, Apple, Meta, Nvidia, OpenAI and Anthropic.
2. The start-up scene draws on this base, and Zurich led the Swiss cantons in venture financing in 2025, although growth capital remains scarce and the market fell back in the first half of 2026.
3. The talent base is highly international and well paid, yet it is constrained by a vacancy rate of about 0.1 per cent in the city of Zurich and by annual quotas for workers from third countries.
4. FINMA addresses AI through its existing framework, with Guidance 08/2024 setting expectations on governance and risk management, while the outsourcing, operational resilience, data protection and cyber reporting rules apply to AI without naming it.
5. Switzerland’s own AI legislation, based on the Council of Europe Convention and sectoral amendments, is still in preparation, so that the EU AI Act and professional rules such as banking secrecy and Article 321 SCC currently define the practical limits for AI in Zurich’s banks, insurers and law firms.
For general information only. This article does not constitute legal or tax advice, does not replace advice on an individual case and creates no lawyer-client relationship. It reflects the legal position as of September 2025.